Scribd
Upload
402 views3 pages

Math 202 - Homework #6, 14.7 Solutions: N N N N

This document provides solutions to homework problems from a math class on field theory. It includes proofs of several statements about field extensions and Galois groups. The document proves that a field extension depends only on the ratio of the elements being taken, determines the Galois group of a cyclotomic field extension, and expresses the generator of the unique quadratic subfield of a cyclotomic field in terms of p.

Uploaded by

MorvaridYi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
402 views3 pages

Math 202 - Homework #6, 14.7 Solutions: N N N N

This document provides solutions to homework problems from a math class on field theory. It includes proofs of several statements about field extensions and Galois groups. The document proves that a field extension depends only on the ratio of the elements being taken, determines the Galois group of a cyclotomic field extension, and expresses the generator of the unique quadratic subfield of a cyclotomic field in terms of p.

Uploaded by

MorvaridYi
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Math 202 Homework #6, 14.

7 Solutions
May 19, 2015

Sec 14.7, #3. If is a square in F , then F ( ) = F ( ) = F if and only if


is a square inF . So assume is not a square in F , in particular 6= 0. Claim:
only if /
(F )2 . One direction is clear: if = x2
F ( ) = F ( ) if and

) = F (x ) = F ( ). For the converse, suppose


that
with
, then F (

x F
F ( ) = F ( ). Then = x + y for some x, y F , so = x2 + 2xy + y 2 .
Therefore
2xy = 0. Since char(F ) 6= 2, this implies x = 0 or y = 0. But y = 0 implies

= x F , hence is
a square in
F ; as we saw in the beginning of the proof, if
is a square in F then F ( ) = F ( ) = F if and only if is also a square in F (in
2
which
case
were left with the case x = 0, which implies that
/ (F )2 ). Hence

= y , so / = y (F )2 as claimed.
p

Applying
this
result
to
F
=
Q(
2),
we
see
that
Q(
1

2)
=
Q(i,

2) if and only

if (1 2)/(1) = 1 + 2 is a square in Q( 2). Set (x + y 2)2 = 1 + 2


with x, y Q. We obtain x2 + 2y 2 = 1 and 2xy = 1. The first equality is clearly
impossible
to satisfy, since x2 + 2y 2 > 0 for any real numbers x and y. Therefore
p

Q( 1 2) 6= Q(i, 2).

#7. (a) We have ( n a)n = ( n a ) = (a) = a, hence ( n a) = n a for some nth


rootof unity .
Since F , iterating this equality m times for any integer m yields
m n
m n
( a) =
a. Therefore the order of is exactly equal to the multiplicative order
of . Since has order d, we see that must be a primitive dth root of unity.


(b) By part a, both ( n a)/ n a and ( n b)/ n b are primitive dth roots of unity. Hence
the first of these can be written as the second to the power i for some integer i rela
i

tively prime to d. The equation ( n a)/ n a = (( n b)/ n b)i yields ( n a)/( n b ) =


i

i
n
a/ n b . Therefore n a/ n b is fixed by Gal(K/F ) = hi and hence lies in F .
i

(c) Again
one direction is easy: If a = bi cn1
and b = aj cn2 then n a = c1 n b F (n b)

and n b = c2 n a F ( n a), so F ( n a) = F ( n b). For the converse, if F ( n a) = F ( n b),


then part (b) implies that a = bi cn1 for some c1 F and by symmetry b = aj cn2 for
some j and some c2 F as well. This completes the proof.
#9. Since this problem relies on previous exercises (namely, #21 and #26 of Section 2),
let us write out what we need from those exercises along the way. Write Gal(K/F ) =
1

hi. The linear independence of characters implies that


TrK/F = 1 + + 2 + + p1 6= 0
as a function K K. Therefore, there exists an element K such that TrK/F 6= 0
Let
1
=
(() + 2 2 () + (p 1) p1 ()).
TrK/F
Then () is equal to
1

 2

( () + 2 3 () + (p 1) p ()) (() + 2 2 () + (p 1) p1 ())

TrK/F
1
=
( () 2 () 3 () p1 ()) = 1,
TrK/F

so () = 1. In particular, since () 6= , we see that 6 F and hence


F () = K (since [K : F ] = p is prime, the only subfields of K are F and K itself).
Now let a = p . Note that
(a) = ()p () = ( 1)p ( 1) = p 1 ( 1) = p = a.
(Here we used the freshmans binomial theorem, (1)p = p 1 in characteristic p.)
Since generates Gal(K/F ), we find that a F . Therefore satisfies the polynomial
xp x a F [x], and since this polynomial has degree p, it is the minimal polynomial
of .
Note that the roots of this polynomial are + k for k Fp , and the element i
Gal(K/F ) acts on the roots by adding i.
P
#11. (a) Since H (Z/pZ) is the subgroup of squares, it follows that 0 = a square pa
P
and 1 = b non-square pb . A generator of (Z/pZ) is a non-square. So Gal(Q(p )/Q)
satisfies (p ) = pg for a non-square g (Z/pZ) . Now the product of a square and
non-square is a non-square, and the product of two non-squares is a square (this is
just
law in the group (Z/pZ) /H of order 2. We therefore have (0 ) =
P the group
P
ag
b
a square p =
b non-square p = 1 and similarly (1 ) = 0 .
(b) If e is a generator of (Z/pZ) , then the squares are precisely the powers e0 , e2 , . . . , ep3
and the non-squares are precisely the powers e1 , e3 , . . . , . . . ep2 . Then
0 + 1 =

p2
X

pe =

k=0

and
0 1 =

p2
X

k
(1)k pe

p2
X

k (p ) = (p , 1)

k=0

k=0

p2
X
k=0

(1)k k (p ) = (p , 1)

by definition. Finally, it is clear that (p , 1) =


polynomial p (x) = xp1 + xp2 + + 1.

Pp1
j=1

pj = 1 since p satisfies the

(c) As i ranges from 0 to p 1, i2 modulo p covers the value 0 once and each of the
squares in (Z/pZ) twice. Therefore
g = 1 + 20 = 1 + 0 + 0 = 1 + 0 + (1 1 ) = 0 1 = (p , 1).
(d) Since g = 0 1 , and H fixes 0 and 1 whereas G H swaps 0 and 1 , it is clear
that (g) = g if H and (g) = g if 6 H. In particular, g lies in the fixed field
of H, which has degree 2 over Q since H is an index 2 subgroup of G, and g 6 Q since
g is not fixed by all of G. Therefore [Q(g) : Q] = 2. Also, since complex conjugation
is equal to 1 , our description above shows that g = g with the sign depending on
whether 1 H, i.e. whether or not 1 is a square mod p.
(e) The problem is mostly solved in the parenthetical comment, which we dont copy
k
over. For k = 0 to p2, consider k (p )/p . If = e as above, then k (p )/p = pe 1 .
Pp2 j
(1) = p 1.
If k = 0, then clearly this is just 1. In this case we obtain j=0
k

But if 0 < k < p 1, then ek 6 1 (mod p) since e is a generator of (Z/pZ) , so pe 1 is


P
j
a primitive pth root of unity . In this case we obtain p2
j=0 () = TrQ(p )/Q = 1.
An explanation for this last equality is that TrQ(p )/Q is the negative of the coefficient
of the second highest exponent term in the minimal polynomial for , which is p (x) =
xp1 + xp2 + . . . + 1. Putting this together with the computation in the hint we obtain
gg = (p 1) +

p2
X

(1)k = (p 1) + 1 = p.

k=1

(f) If p 1 (mod 4), then 1 is a square mod p, so g = g and hence g 2 = gg = p. If


p 3 (mod 4), then 1 is not a square mod p, so g = g and hence g 2 = gg = p.
2
(p1)/2
These can be combined by writing
p since (1)(p1)/2p= 1 depending
p g = (1)
on p mod 4. Therefore g = (1)(p1)/2 p and it follows that Q( (1)(p1)/2 p) is
the unique quadratic subfield of Q(p ).
Final note: Figuring out exactly what sign occurs here in the equation
q
g = (1)(p1)/2 p
is a problem that Gauss spend a lot of effort thinking about.

You might also like